Bio-Medical Materials and Engineering 26 (2015) S583–S592 DOI 10.3233/BME-151349 IOS Press
S583
Preliminary study on verifying the detection of gait intention based on knee joint anterior displacement of gait slopes Changho Yua, Seung Rok Kang a, Giltae Yang a, Chul Un Hong a, Hyung Jong Leeb, Do Young Ohc and Tae Kyu Kwona,* a
Division of Biomedical Engineering, Chonbuk National University, Jeonju-si, Jeonbuk, Korea Department of Healthcare Engineering, Chonbuk National University, Jeonju-si, Jeonbuk, Korea c EasyMove Co., Ltd., 230 Anyang-ro, Anyang-si, Gyeonggi-do, Korea b
Abstract. This study investigated the feasibility of the Infrared (IR) sensor-based walking aids for detecting the gait intention. To compensate for the defects of Force Sensing Resistors (FSRs) or force sensors, such as the velocity control problem on gait slopes, we used IR sensors to investigate knee joint anterior displacement in order to recognize the gait intention. We also measure leg muscle activities and foot pressure, in order to verify our investigation. We placed two IR sensors on the rollator center to sense left and right leg walking intentions. We took EMG signals of four leg muscles, and analyzed them. Foot pressure analysis parameters were the measured force and mean pressure. We conducted experiments on twenty young healthy adults. The results show that knee joint anterior displacement increases according to gait slope and velocity. We confirm similar results of knee joint anterior displacement through the IR sensors. Keywords: Rollator, foot pressure, EMG, IR sensor
1. Introduction Human aging is characterized by a progressive decline in the ability to perform basic mobility tasks eventually resulting in functional dependence and disability. The muscle power of the elderly tends to significantly deteriorate after the age of 65, due to aging [1, 2]. Weakened leg muscles make it difficult for pedestrians to walk properly. Most elderly people with weakened leg muscle use sticks or walking aids in their daily lives [3-6]. In order for the elderly to conveniently and happily lead their daily lives, they should be free from difficulty in walking as the most basic activity. Mobility assisting devices including electronic wheelchairs are the technology needed in the field of welfare instruments as the groups of handicapped and the elderly increase, and the marketability of these devices is increasing. In order to solve issues of convenience and improvement of functions, the necessity for objective and generalized International Organization for Standardization (ISO) regulations, as well as the design required by users, and engineering/technological aspects is increasing [7-9]. *
Address for correspondence: Tae Kyu Kwon, Department of Biomedical Engineering, Chonbuk National University, Jeonju-si, Jeonbuk, Korea. Tel.: +82-63-270-4066; Fax: +82-63-270-2247; E-mail:
[email protected]. 0959-2989/15/$35.00 © 2015 – IOS Press and the authors. This article is published with Open Access and distributed under the terms of the Creative Commons Attribution and Non-Commercial License.
S584
C. Yu et al. / Preliminary study on verifying the detection of gait intention based on knee joint anterior displacement
Intelligent robot type walking aids that are developed to support the lives of the handicapped and the elderly basically support the elderly to walk properly, help them sit down and stand up, and let the elderly move to a destination safely and conveniently with top priority. Representative studies dealing with walking aids for the handicapped and the elderly have been conducted for technology in developing the sensors designed for seamless control of walking aid vehicles of seniors, vehicle movement technology for stably controlling vehicles, and obstacle-avoiding technology for solving the difficulty of the handicapped and the elderly. Such studies tend to help the handicapped and seniors use walking aids more easily [10-16]. The most likely technology of identifying the gait intention by using walking aids is FSR sensors or force sensors [17, 18]. Such technologies do not cause particular problems for the safety of users when walking on the flat. However, there might be an issue when moving on slopes. In particular, they need to slow down when walking on downward slopes. However, much strength on the hand knob is needed to make walking aids move slowly. Current walking aids might increase the walking speed, causing the risk of accident to increase [19]. In addition, according to one particular study dealing with the movement of lower bodies, two IR sensors attached underneath the walking aids estimated the movement of shins when walking and identified the gait intention. However, there has been a lack of study dealing with gradient conditions [20]. Therefore, this study uses IR sensor-based walking aids to measure the forward displacement of knee joints depending on the gradient conditions when walking, to identify the gait intention. In addition, we verified the results of identifying the gait intention from IR sensors by the result of evaluation of electromyogram and foot pressure. 2. Experimental methods 2.1. Subjects Thirty healthy persons were enrolled in the study on a volunteer basis. Twenty volunteers without medical history from injury of exercise within 6 months, and also without experience of exercise within the recent 3 months were randomized in to this experiment (20 males: age 24.3±1.5 years; weight 75.6±4.6 kg; height 174.7±5.3 cm). Prior to the beginning of the examination, conforming to the Declaration of Helsinki (1964), written informed consent was obtained from all subjects.
Fig. 1. Illustrations of knee joint anterior displacement using IR sensor: (a) Top view, (b) Side view.
C. Yu et al. / Preliminary study on verifying the detection of gait intention based on knee joint anterior displacement
S585
2.2. Systtem configuration Figuree 1 shows the evaluation of forward displacemen d nt of the left and a right kneee joints by installing two IR sensors s (GP2 2Y0A02YK0 0F, SHARP Co., C Japan) in i the middlee of the tablee of an active walker. The activ ve walking aid a used in the t experimeent, Active Walker, W is a product p that passes the criteria c of KPS perrformed by the Ministry y of Comm merce, Industtry, and Eneergy, is com mprised of ultra-light u aluminum m materials, and weigh hs 8.9 kg. Depending D on o the height of users,, the chair height is adjustable from 470 to 620 mm, and the hand d knob heigh ht is adjustab ble from 7700 to 1000 mm m. The IR sensor iss an infrared light distancce measurem ment sensor th hat allows th he distance too be adjusted d from 20 to 150 cm m for estimaation of the knee k joints. The T operation n voltage is analogue a outtput from 4.5 5 to 5.5 V, and the estimated e sp peed is 38.3 ± 9.6 m/s. Fo or defining the t forward displacemen d nt of knee joiint in this study, we w measured the value of o difference on peak vaalues marked d on the IR sensor when n moving forward or backward ds on the kneees when wallking to iden ntify the gait intention.
Fig. 2. (a)) Experiment piicture for measuurement of IR sensors, s (b) Exp periment picturee for measurem ment of electrom myography.
Fig. 3. (aa) Experiment picture p for meassurement of fooot pressure, (b) Algorithm flow w for determinattion of the gait intention.
S586
C. Yu et al. / Preliminary study on verifying the detection of gait intention based on knee joint anterior displacement
For the measurement of forward displacement of the knee joints, we established a baseline for estimation of the IR sensor after attaching markers to both knees. For real-time data transmission, a DAQ board (USB-6009, National Instruments Co., USA) was installed underneath the chair of the walking aid in connection with a computer. We used NI Labview 2011 (National Instruments Co., USA) system to design the software to identify the gait intention. When the subject starts walking, data of the forward displacement of knee joints of the user were saved, to enable design of a system for identifying walking will through analysis. IR sensor data were estimated for a minute while walking. Figure 2(a) shows the IR sensor evaluated the forward displacement of the knee joint. Measurement was performed on a treadmill (Gait Trainer, Biodex Ltd., USA) to control the walking speed and gradient conditions. The slope was adjusted (front, back) by using the treadmill. The slope was divided into front slope, rear slope, and flat ground, and the front/rear slopes were 10 degree, while the flat ground was 0 degree. We set the walking speeds to be 1 km/h, 2 km/h, and 3 km/h. We evaluated the muscle activation and foot pressure using IR sensor-based walking aids as shown in Figures 2(b) and 3(a), to compare and analyze physiological signals, and the result derived from the gradient conditions and walking speed when walking by using the IR sensor-based walking aids. The estimated lower limbs muscles were biceps femoris, rectus femoris, gastrocnemius, and tibialis anterior using the Bagnoli-EMG 8ch with the sampling rate of 1000. In order to remove noise by motion artifact we used a band pass filter from 25 to 450 Hz and which used sterilization alcohol to reduce the skin resistance against the surface. The standardized EMG signals were conducted for the frequency analysis walking the treadmill the probability density function was earned to calculate the spectral energy to know the muscular activity. Root Mean Square (RMS) analysis was applied to analyze the EMG data for estimating amount of work in muscles during gait. For foot pressure, we used a Pedar-X system to estimate the fore/mid/hind-feet. We used force and mean pressure as the analysis parameters. Each insole contained 99 force sensors and a working dynamic range of 15–600 kPa, which were all calibrated using a standard calibration device. The Pedar-X data acquisition software was used to collect and filter data and sample frequency was set at 100 Hz. Subjects wore the same shoes for the estimation, in order to obtain accurate date value on the foot pressure. Figure 3(b) indicates the algorithm flow to identify the gait intention. We analyzed the forward displacement of the knee joints value, the muscle activation value, and the foot pressure result obtained from the IR sensors to identify the gait intention, depending on the gradient conditions and walking speed. We performed the experiment for 30 minutes each time for three weeks. In order to remove error due to different postures between subjects, subjects maintained a consistent posture. In addition, all the subjects were well informed of the objectives and method of the experiment prior to it. 2.3. Statistical analysis We intended this study to verify the validity of identifying the gait intention based on the results of IR sensor values for each walking speed and gradient condition, muscle activation, and foot pressure. We analyzed the estimated data for IR sensor value, muscle activation, and foot pressure by using SPSS 18.0 (SPSS Inc., Chicago, IL, USA). We calculated the average and standard deviation of IR sensor values, muscle activation, and foot pressure values for each gradient term and walking speed,
C. Yu et al. / Preliminary study on verifying the detection of gait intention based on knee joint anterior displacement
S587
Fig. 4. (a) Left knee joint anterior displacement for each slope direction and velocity (*p
